A segmentation analysis: the case of photovoltaic in …...analysis into the adoption and non-adoption of solar PV in Dutch households. It is based on a survey under 817 households

ORIGINAL ARTICLE

A segmentation analysis: the case of photovoltaicin the Netherlands

Véronique Vasseur & René Kemp

Received: 19 December 2013 /Accepted: 19 February 2015 /Published online: 14 March 2015# The Author(s) 2015. This article is published with open access at Springerlink.com

Abstract The paper presents the results of a surveyanalysis into the adoption and non-adoption of solar PVin Dutch households. It is based on a survey under 817households undertaken in 2012. Households are aggre-gated into 4 groups based on whether the adoption isvoluntary or involuntary (when people buy a house withsolar panels) and whether or not the household can beconsidered a potential adopter or rejecter. For these fourgroups, we study and compare the characteristics ofadopters and nonadopters of solar PV. Non-adopters arebroken down in two groups: rejecters and potentialadopters. The segmentation analysis gives more specificinsight in the adoption of PV but can also be used toinsight in the adoption of other technologies and/or inother countries.

Keywords Technology adoption . Photovoltaic (PV) .

The Netherlands

Introduction

Solar energy systems, i.e., photovoltaic (PV), continueto gain attention in the Netherlands as consumers seekalternatives to increasingly expensive conventional en-ergy sources. Concerns about energy usage and energy

costs are expected to further the consumer demand forPV, accompanied by a rapid expansion in the acceptanceof these systems in the future years. The adoption of PVis driven by this consumer demand and is characterizedby the number of individuals or households that decideto adopt or reject this technology. We are interested inthe pioneers who have adopted a PV system, which kindof people adopt it and what are for example their demo-graphic characteristics. But, also the people whorejected a PV system are of importance. What kind ofpeople are they? Given the current activity and interestin solar energy and the future growth expected in thisindustry, it is important that we are able to identify thesepeople to create insight in the adoption process. Littlescientific research is presently known concerning theindividuals adopting a PV system.

To our knowledge, there are some studies which in-vestigate how and why consumers may transition towardadopting sustainable energy technologies. Most interest-ing is the study of Axsen et al. (2012), who grouped theirsample into the greens that are Bengaged,^ those who areBaspiring^ and those who are Blow-tech,^ and the non-environmental groups, whereby it is distinguished be-tween the Btraditional^ and the Btechies.^ However, inour opinion, it highly depends on the technology inquestion whether people are willing to adopt, even withina group Bgreen-engaged^ it is impossible to predict thedecision. A study by Pedersen (2000) showed that eveninconsistency in the purchase of various green productsand/or technologies exists. He contends that the purchaseof a green product cannot be predicted based on thepurchase of another green product (Faiers and Neame

Energy Efficiency (2015) 8:1105–1123DOI 10.1007/s12053-015-9340-8

V. Vasseur (*) : R. KempInternational Centre for Integrated Assessment andSustainable Development, University Maastricht,P.O. Box 616, 6200 MD Maastricht, The Netherlandse-mail: [email protected]

2006). Thus, it is not because, for example, buying anelectric vehicle that one will also buy solar panels forexample. The opposite is obviously also true: It is notbecause a polluted product (e.g., a plane trip) that onedoes not buy green products (e.g., solar panels).

Drawing on the theory of market segmentation, weintroduce a typology which can be used for analyzing theadoption of technological innovations, in particular PV.Different typologies to classify people exist. Some re-searchers prefer to highlight the connectedness of differ-ent subtypologies within a typology; then, the term seg-mentation instead of typology is often used. Segmentationis defined as a process of dividing one population intosmaller subpopulations (i.e., segments or groups), whichare characterized by different needs, characteristics or be-haviors, including their response to the way they areapproached and affected (Geest et al. 2008). In this re-search, we are interested in the different ways of thinking,beliefs, and perception of people which make the conceptof segments more practical as it highlights the interconnec-tedness of parts related to a larger populace (the Dutchpopulation). So, the objective of this study is to introduce asegmentation model which allows us to answer the ques-tion whether adopters and rejecters of a PV system con-sider the same or different values/attributes. Moreover, theresearch method is accessible and workable for otherresearchers who such to gain insight in adoption processes.

Section 2 offers a theoretical background, we discussedthe theory of market segmentation. Section 3 offers theresearch method and data collection. Based on the gainedinsights in segmentation literature, we introduce a newmodel in section 4 which can be used for analyzing thediffusion and adoption of PV in the Netherlands. Theoverall purpose of this section is to determine empiricallyhow the groups differ from each other. We take the demo-graphic characteristics, geographical characteristics, phys-iographical characteristics, and behavioral characteristicsinto account. Section 5 offers a reflection of our segmen-tation model, and we analyze how the model met differentcriteria for good segmentation. Finally, section 6 provides adiscussion and conclusion.

Theoretical background

Technology adoption cycle

Consumers have different personal characteristics andtraits and do not all adopt innovations (a new technology)

at the same time (Beal and Bohlen 1957). Beal et al.(1957) divide the diffusion of new ideas into five stages:awareness, interest, evaluation, trial, and adoption.Interesting in Beal and Bohlen’s discussion of these fivestages is how the most common way for people to learnabout new technologies change at each step in this pro-cess. When it comes to individuals, Beal et al. (1957)introduced a technology adoption lifecycle which dividedpeople into categories that are determined by how soonthey adopt new technologies. This is where they dividedpeople into the categories of innovators, early adopters,early majority, (late) majority, and laggards or non-adopters. The innovators, early adopters or early majorityare individuals or firms investing at an early stage of thediffusion of new technologies. They have a large net-work, access to information, investment capital, an edu-cational level, or experience above average. The (late)majority and non-adopters or laggards are peopleinvesting on a later stage, they are older than peopleinvesting at an early stage of the innovation diffusionand they have a smaller network and are less educated.See Table 1 for a more elaborated description of thedifferent categories.

The technology adoption cycle should be viewed as arelative concept. It seems to be straightforward that thegroup of innovators depends on the technology in ques-tion, and it does hardly account for differences in thecircumstances of users and difference of the preferences.

Market segmentation

Market segmentation is a marketing strategy whichrefers to the process of dividing a potential market intodistinct subsets of consumers who have common needsand priorities and selecting one or more segments as atarget market to be reached with a distinct marketingmix. The basic idea behind market segmentation is thatconsumers are not all the same and that one strategy willnot work for all customers. This has led to break downlarge markets into smaller segments, each of which ismore narrowly defined than the overall target market.Different authors have discussed market segmentation,e.g., Abell and Hammond (1979), Gankema and Wedel(1992), Hessing and Reuling (2003), and Schiffman andKanuk (2010). It is a common view that a good seg-mentation has to comply with seven criteria (Abell andHammond 1979; Gankema and Wedel 1992; Hessingand Reuling 2003; NetMBA 2010).

1106 Energy Efficiency (2015) 8:1105–1123

– Identifiable. It should be clear to which segmentsomebody belongs.

– Accessible. People should have the opportunity tomove to another segment.

– Substantial (size). The segments should be suffi-ciently large and there should not be too muchgroups consisting of only a few people.

– Unique needs/heterogeneous. The segments shoulddiffer clearly and with clear differences between thesegments.

– Stable (durable). The segments should be relativelystable (minimize often and/or easily changes be-tween segments).

– Homogeneous response. Members within a seg-ment should react in a comparable way to arousals,e.g., advertising and information.

– Conducive to steering/affecting behavior (influen-tial). The typology should offer ideas on how hu-man behavior within each segment can be steered.

With these criteria in mind, there are many conceiv-able ways in which a market can be segmented. Thinkabout where consumers are located, who is purchasing aproduct or service, why consumers buy what they do,and so on. However, the segmentation itself is based ona limited number of characteristics. Four commonlyseen characteristics are found in the market literature1,e.g., Beane and Ennis (1987), Wedel and Kamakura(2000), Doornbos (2004), and Schiffman and Kanuk(2010). Demographic characteristics2 refer to age, gen-der, family composition, education level, housing type,and income. Understanding who consumers are (middleaged woman, high educated) will enable you to moreclosely identify and understand their needs, product, andservices usage rates and wants. Geographical character-istics divide the total potential market into smaller sub-groups on the basis of geographic variables (e.g., city,region, province, postal code, and population density.Psychographic characteristics, or lifestyle characteris-tics, refer to activities, interests, opinions, attitude, andvalues. Segmenting consumers into lifestyles is basedon the notion that a person’s lifestyle has a direct impacton their interests in products and services. Behavioralcharacteristics divide consumers into groups accordingto their motive to buy/benefits sought (price, esthetic,functionality, idiosyncratic preferences), readiness tobuy, and occasions (event that stimulate the purchase).

Other consumer-rooted characteristics used to seg-ment markets are personality traits and socioculturalvalues and beliefs. The key consumption-specific seg-mentation factors are usage behavior (including usagerate and situation), benefit segmentation, and brandloyalty and relationship (Schiffman and Kanuk 2010).

The most common category used in market segmen-tation is demographics. However, this group does notprobe into why consumers buy what they do; hence, itdoes not offer an understanding of what motivates con-sumers to buy certain services or what types of person-alities favor a product or brand over another. Answerson these question combined with demographic data areeven more valuable. That is also the reason why a singlecharacteristic is almost never used alone. Practically, all

Table 1 Categories of technology adoption cycle

Name Description

Innovators—2.5 % First, to adopt in a very early stage of theinnovation process. They are willing totake risks, often have substantialfinancial resources and a technicalknowledge

Earlyadopters—13.5 %

Role model for other members of the socialsystem. They are aware of theirimportant position and try to maintainthis position by making quick judiciousdecisions which will trigger the mass toadopt an innovation

Earlymajority—34 %

Adopts a new technology when they seethat the implementation was successfulin the early adopters group. This grouptakes its time to make a deliberatedecision in order to avoid the start-upproblems of an innovation

Latemajority—34 %

Adopts an innovation when there is apressure from the environment or whenthe innovation has proven higherperformance. This group has a lowersocioeconomic status

Laggards/non-adopters—16 %

Last to adopt an innovation, for whotechnology is unattractive. They are veryconservative, isolated from the rest of thesocial system, and often have limitedresources

From Beal et al. 1957; Rogers 2003

1 It is important to note that sometimes, textbooks classify thecharacteristics differently. For example, we integrate Bbenefitssought^ as being a Bbehavioral characteristic,^ while some text-books reported that they should be separated out. And, sometextbooks will list geo-demographics—a combination of geo-graphic and demographic measures—as a separate category.2 Socioeconomic characteristics are covered by this group.

Energy Efficiency (2015) 8:1105–1123 1107

segmentation models are in the form of hybrid segmen-tation. Hybrid segmentation, also referred to as multi-variate segmentation, refers to using multiple segmen-tation characteristics in order to determine the marketsegment (Schiffman and Kanuk 2010). Below someprimary examples of hybrid segmentation models aredescribed. Important to keep in mind, it is not ourpurpose to be exhaustive but rather to broaden insightin combining different segmentation categories.

Geo-demographic segmentation—PRIZM

Geo-demographic segmentation involves a combinationof geographic and demographic factors. This segmenta-tion is based on the notion that people who live close toone another are likely to have similar financial means,tastes, preferences, lifestyles, and consumption habits(Schiffman and Kanuk 2010). Jonathan Robbin, thefounder of Claritas Inc., is seen as the father of geo-demographic segmentation and introduced in 1974 thePRIZM (Potential Rating Index for ZIP Markets) seg-mentation system. PRIZM divides the US populationinto 15 social groups (e.g., Landed Gentry, 2nd CityCenters, Urban Cores) based on differences in socioeco-nomic status and urbanization. Those 15 groups arefurther subdivided by measurable differences in house-hold composition (e.g., family structure), mobility, eth-nicity, and housing. It defines every US household interms of 66 segments to help marketers discern thoseconsumers’ likes, dislikes, lifestyles, and purchase be-haviors (Heitgerd and Lee 2003; Wedel and Kamakura2000). The segment names are meant to catch interestand convey a general sense of the character of an area;however, they should not be interpreted literally.

Psychographic/lifestyle segmentations—VALS and LOV

One of the most widely popularized approaches to life-style research for market segmentation is the values andlifestyles (VALS) segmentation model developed atStanford Research Institute (SRI) by Mitchell (1983),drawing on the theoretical base of Maslow’s (1954)needs hierarchy and the concept of social character(Reisman et al. 1950). The VALS model is developedto determine different classes of people who had varyingvalues, attitudes, and lifestyle. This typology classifiesthe American adult population into eight distinct sub-groups based on a specific set of attitudinal and demo-graphic questions that drive consumer behavior. The

model illustrates two critical concepts for understandingconsumers: primary motivation and resources (e.g., in-come, education). The combination of motivations andresources determines how a person will express himselfor herself in the marketplace as a consumer (Schiffmanand Kanuk 2010; Kahle et al. 1986; Wedel andKamakura 2000). The eight VALS segments are inno-vators, thinkers, believers, achievers, strikers,experiencers, makers, and survivors.

One alternative to VALS is the list of values (LOV)which was developed by researchers at the University ofMichigan Survey Research Centre (Kahle et al. 1986).LOVwas developed from a theoretical base of Feather’s(1975), Maslow’s (1954), and Rokeach (1973) work onvalues in order to assess adaptation to various rolesthrough value fulfillment. A list of nine values is usedto classify consumers, including self-respect, security,warm relationships with others, sense of accomplish-ment, self-fulfilment, sense of belonging, being wellrespected, fun and enjoyment in life, and excitement.These values are used to classify people on Maslow’shierarchy, and they relate more closely to the values oflife’s major roles than the values in Rokeach ValueSurvey (ordered value system). Respondents have beenasked to identify their two most important values or torank the values (latent value system) (Kahle et al. 1986;Schiffman and Kanuk 2010).

Some similarities are found between both segmenta-tion models: (1) VALS segmentation of achievers andthe LOV segmentation of sense of accomplishment, and(2) VALS classification of believers and the LOV clas-sification of sense of belonging. Whereas in VALS, theindividual is viewed as going from worse to better,within the LOV framework, no such expectation exists.Comparisons of the VALS and LOV model ofsegmenting consumer markets have indicated that theLOV method has certain advantages over VALS, andthe LOV has been found to predict consumer behaviormore often than VALS across a range of brands (Kahleet al. 1986).

MindBase

The segmentation methodology, MindBase, is createdby Yankelovich researchers who examined four years ofin-depth data on American values and attitudes from itsannual monitor surveys comprehensive study ofAmerican opinions on topics such as government,health, sex, business, and religion, which it has carried


out since 1971. From this data, Yankelovich identifiedeight major consumer groups with shared life attitudesand motivations: up and comers, aspiring achievers,realists, new traditionalists, family centereds, individu-alists, renaissance masters, and maintainers. These eightgroups were further divided into 32 distinct subseg-ments for greater differentiation and clarification(Hawkins et al. 2007).

Commercial segmentation models

In the Netherlands, six major research centers have eachdeveloped a segmentation model based on personalvalues and lifestyle. The different segmentation modelsare valuebox model of NFO-TrendBox, metality modelof motivaction, mosaic model of Experian, win modelof TNS/NIPO, censydiam model of synovate, and BSRmodel of SmartAgent Company. Every segmentationmodel is already used in a case related to the buildingmarket, e.g., municipality Almere by Experian. Theseresearch centers used the models to help their customerswith the segmentation. In the past, a disadvantage ofthese models was that there was almost no connectionpossible to reach segments of the market. In recentyears, research centers have coupled their segmentationmodel with large address files or large self-made data-bases. In this way, the translation of the results to reachpotential customers is easier. However, the researchmethods are not communicated since this is confiden-tial; therefore, we include in the underlying idea of thesemodels as an appendix (see Appendix).

To sum up, market segmentation is all about identi-fying specific groups of people based on common char-acteristics. There are virtually dozens of ways that amarket might be segmented and the segments chosenwill depend on the products or services it offers. It isimportant to keep in mind that with the identified seg-ments, one will decide which strategy is best for a givenproduct or service, and now and then, the best optionarises from using different strategies in conjunction.

Research method and data collection

With the theoretical background of market segmentationin mind, it is intended to introduce a segmentationmodel which can be used for analyzing the adoption oftechnological innovations.We focus on PVas it is one ofthe most promising low carbon energy sources. While

the worldwide application of PV is growing fast, theNetherlands is lagging behind which clearly constitutesa case of slow diffusion. By studying this case, we findout which kinds of people use the technology alreadyand which kind of people rejects the technology.

Empirically, our research is based on original data onthe perception on solar energy in the Netherlands col-lected via an internet questionnaire. The data gatheringtook place in September 2011, and the responseconsisted of 817 completed and usable questionnaires.The data is used to determine empirically a new seg-mentation model for technological innovations.

After some trial and error, we figured out that it isimpossible to group our sample into different lifestylegroups based on VALS, PRIZM, or the like. For that, amuch bigger sample and much more questions on atti-tudes and values would be required. But, the strength ofthe survey that could be exploited in more detail is that itmeasures the environmental attitudes and concerns ofadoption/non-adoption of PV of the respondents prettywell. We explored the use of a factor analysis, a clusteranalysis, a principal components analysis, and chi-squared automatic interaction detection (CHAID) anal-ysis. All techniques are used not only (1) to see therelationship between the items in the questionnaire andunderlying dimensions but also (2) to reduce a larger setof variables to a smaller set of variables that explain theimportant dimensions of variability. A factor analysisaims to find underlying latent factors3, whereas princi-pal components analysis aims to summarize observedvariability by a smaller number of components. Clusteranalysis divides people into groups that are meaningfuland interesting; however, these groups are not associat-ed either with an outcome measure such as likelihood ofpurchasing a product or with background data thatwould allow them to be identified so that specific mes-sages can be addressed to them. Finally, CHAID iden-tifies segments that are related to an objective and easilyidentifiable. However, this method does not work verywell with attitudinal data, as the segments tend to lackdepth and complexity, or their complexity is uninterpret-able. As such, we observed that these analyses are notuseful for the aim of our research, as we will gainspecific insights into the adoption and non-adoption ofsolar PV in Dutch households. Therefore, we decided to

3 Doing the factor analysis to group the different behaviours andthereby increase the reliability is very difficult given that correla-tions are fairly low between behaviours.


cluster some specific questions in order to reduce thelarge set of items into a smaller number of dimensions.

In this research, we propose a hybrid segmentationmodel which we introduce in the next section to analyzewhether PV users are identifiable as small number ofrelatively homogeneous groups of technological users,based on their adoption or rejection of a specific tech-nological innovation. The introduced model is testedwith the criteria of o.a. Gankema andWedel (introducedin BTheoretical background^ section), namely identifi-able, accessible, size, unique needs (heterogeneous),stable, homogeneous response, and conducive ofsteering. Our focus is placed on demographic, geo-graphic, and psychographic characteristics and motiva-tions (behavior characteristics) rather than on feelings orintentions. We are interested in the personal characteris-tics of Dutch citizens in relation to sustainable energysources and in particular PV. In order to gain insight inPV users, the following dimensions are important.

– Demographic characteristics. In our questionnaires,we included questions about age, possibly income,education, and gender to analyze if some character-istics occur more often in a specific group.

– Geographic characteristics include the housingtype, housing situated, ownership, and number ofresidents per dwelling as these characteristics influ-encewhere PV systems appear. The domestic sectorin the Netherlands is divided over three types ofownership. Each represents a different type of de-cision maker with respect to the purchase of PV: (1)owner-occupied sector in which the residents them-selves are the decision makers, (2) private rentalsector in which private landlords make the invest-ment decision, (3) public rental sector in whichhousing associations make the investment decision.Broadly five types of houses can be distinguished inthe Dutch domestic sector: (1) detached (free stand-ing), (2) middle of a row, (3) semi-detached, (4)apartment, and (5) farms. Furthermore, a house canbe situated in a city, village, or countryside.

– Psychographic characteristics include activities,opinions, and values in our questionnaire.Activities give an insight into attitudes, norms,and values of people. More specifically, it attemptsto predict specific buying habits and preferences ofconsumers. Often, there is a discrepancy betweenwhat people say they wish to do and their actualbehavior. Activities used in this research are

recycling of paper, avoidance of unaddressed ad-vertising, energy efficient equipment (A-label), andavoidance of car use and water conservation. Weasked the respondents to indicate their contributionwith regard to sustainability within their own life-style with a number of indicators using 4-pointLikert scale ranging from (almost) always to never.The respondents which answer the question with(almost) always or regularly are seen as people whobehave sustainable on that question. Opinions andattitude include the reliance on other people’s ad-vice and approval. To what extent people makedecisions alone or dependent of others, have neigh-bors, family, or friend, is an important role in deci-sion making and/or in their behavior. It also in-cludes to what extent behavior is determined byhabits that exist for decades. Another characteristicis being traditional or modern. Traditional meansthat people conform themselves to habits, rules, andexpectations from a group. The opposite, modern,refers to societies in which not a lot of habits, rules,and expectations exist. Nineteen questions regardpersonal preference were asked. Answering eachquestion implied making a choice between twoopposite possibilities, agree, or disagree. Finally,values can pertain to how a group of individualsfeels about certain characteristics/attributes.Depending of these attributes, people decide toadopt or not adopt a PV system; therefore, it isinteresting to know what people consider as impor-tant product attributes for a PV system and howthese attributes are ranked. Based on focus groupdiscussion4, we proposed that a PV system havefive predominant attributes. These are price, effi-ciency, lifetime, integration, and attractiveness. Weasked the respondents to rank, from very importantto not important, these attributes which are impor-tant in the decision process to adopt or not adopt asystem.

– Behavioral characteristics include the motivationsand the barriers to adopting a PV system. We askedthe respondents who have adopted a PV systemthemselves to rate several aspects in deciding to

4 Within the Organext project, expert interviews with members ofthe project amongst which are R. Kemp (Maastricht University)and J. Manca (Hasselt University) were conducted led by S. Lizin(Hasselt University).


adopt a PV system by importance, maximum threeanswers were possible. There is also a section in thequestionnaire which aimed to find out which as-pects non-adopters of PV find important in order toobserve the barriers for and during adopting a PVsystem. Different product benefits and costs aretaken into account.

A segmentation model for PV in the Netherlands

At the moment, an investment in PV requires a consid-erable run of money. Citizens need an average to aboveaverage income for the purchase of PV. The payback ofPV is around 10 years, and the ideal situation is that thecitizen has this time still in prospect. The dimension lifesituation is not straightforward, and in both situations,you can buy a system. However, the context of beingsingle of having a family can make a difference, forexample, a well-educated single can afford a systemwhile a single earner family man cannot afford the samesystem. Also, the other way around is possible. As acitizen has a home in the private or public rental sector, itis likely that these people do not buy PV panels bythemselves, while for an owner of a house, it can beprofitable. The difference in type of citizen has an influ-ence on the purchase of such innovative technologies; amodern citizen shall easier buy an iPOD than a tradi-tional citizen.

To reduce the large set of items in our questionnaireto a smaller number of dimensions and components(groups), we explored the use of different analysis (seeBResearch method and data collection^). We observedthat these analyses are not useful for the aim of ourresearch as we will gain specific insights into the adop-tion and non-adoption of PV in Dutch households. Thepurchase of a green product cannot be predicted basedon the purchase of another green product; therefore, wedecided to cluster some specific questions in order toreduce the large set of items into a smaller number ofdimensions. In order to understand the meaning of tech-nology for an individual, it is not sufficient to only lookat the abovementioned dimensions (e.g., income, age);more important is to obtain insight into the usage ofsustainable technologies and the sustainable minded-ness of people. Therefore, the attitude of the citizenswe are dealing with is a strong determining dimension inthis research, vertical axis in Fig. 1. This dimension refers

to the attitude citizens have on the technology, positiveversus neutral or negative, while the abovementioned di-mensions are largely influenced by external circumstancesin which the attitude plays no role. Income, life phase, andhome ownership may in itself be decisive factors. Thesecond determining dimension in this research refers tothe decision-making process of major technological inno-vations, see horizontal axis in Fig. 1. An important aspectis the consideration of the adopters and non-adopters.

In this research, the attitude of PV adopters andindividual preferences (adoption or not) is determinedby using different questions. First, we asked whether therespondents have a PV system in their possession. Ifthey have, then we asked who the purchase of thesystem has decided. If the respondent decided the pur-chase by themselves, we label this respondent with apositive attitude; if not, we label this respondent with aneutral or negative attitude. If they have not a PV systemin their possession, a distinction is made between re-spondents who are willing to purchase a system (theyindicated that they are in the orientation phase or thatthey will consider the purchase when more people de-cided to opt for a system) or not; we label these respec-tively as non-adopters with a positive attitude and non-adopters with a negative/neutral attitude.

Taking these dimensions together, we constructed asegmentation model for PV in the Netherlands, seeFig. 1. The number of segments is determined by thenumber of axis used which corresponds with segmenta-tion literature. The figure shows that different attitudesand individual preferences (adoption or not) can bedistinguished in four groups: voluntary adopters, invol-untary adopters, potential adopters, and rejecters.

Better understanding of the adopters (voluntaryand involuntary)5 and non-adopters (potentialadopters and rejecters) allows us to determine howthe groups differ from each other. In this way, wecan compare the choices and considerations for thedifferent adopters. In the section below, the differentadopters are discussed according to the characteristicgroups. An overview of the characteristics is alsogiven.

5 The number of voluntary adopters is very small, and therefore, itmay be quite difficult to make conclusions about the whole pop-ulation; however, the data is suitable to determine how the groupsdiffer from each other.


Demographical characteristics

The age spider diagram (see Fig. 2) suggests that themajority of the voluntary adopters are located in thecategory of 50–59 years, while the involuntary adoptersare concentrated around age 40. Non-adopters appearmore concentrated from categories 40 to 59.

Concerning income, the majority of the respondentshave an income between 15,000 and 36,000 euro peryear. As expected, the respondents with an income lessthan 15,000 euro per year are respectively representedby the group rejecters, involuntary adopters, potentialadopters, and finally voluntary adopters. So, the attitudeof the people with the lowest income is more negative orneutral than the people with an income between 36,000and 60,000 euro per year. This latter group is morerepresented by the respondents who have a positiveattitude for PV, the group voluntary and potentialadopters. These results show that voluntary adopters ofPV have higher income than the average population.This is in line with the results of Labay and Kinnear(Labay and Kinnear 1981), who examines PV within anadoption and diffusion of innovation frameworks in theState of Maine. A case study on the city of Groningenfrom Jager (2006) comes also to similar findings. Thisstudy analyses factors that lead to a faster diffusion ofPV in society from a behavioral perspective.

With regard to education, we found that adopters,especially voluntary adopters, have a higher educationthan the non-adopter. Potential adopters and rejectersappear to be very similar. In gender, the adopter andnon-adopters appear to be also very similar.

Geographical characteristics

In Fig. 3, the spider diagram of the different geograph-ical characteristics is given.

We see that the respondents who have an own househave a more positive attitude than the respondents whorent (public or private). The majority of the home-owneris voluntary or potential adopter. The majority of thegroup of adopters (voluntary and involuntary) lives in avillage, while the group of non-adopters (voluntaryadopters and rejecters) lives in a city. Concerning hous-ing type, the majority of the respondents in every grouplives in a middle of row dwelling (non-detached dwell-ing). But, detached dwellings are even popular for thegroup voluntary adopters, and semi-detached dwellingsare almost even popular for the group involuntaryadopters. The second selected housing type of the grouprejecters is the apartment which can be identified asphysical barrier, as it is assumed that this group ofpeople does not consider the option of solar PV panelsindividually. The last characteristic we discuss is thenumber of residents. The majority of the voluntaryadopters live with two people in a dwelling, while thegroup of involuntary adopters lives with three or fourpeople. Potential adopters and rejecters appear to be alsovery similar.

Psychographic characteristics

The opinion spider diagram (Fig. 4) suggests that cli-mate change is a concern for people with a positiveattitude for PV; thus, the voluntary and potential

Positive

Neutral or

negative

Adopters

Involuntary adopters

Voluntary adopters

Rejecters

Potential adopters

Non adopters

Decision making process

Attitude of PV adopters

Fig. 1 Segmentation model forPV in the Netherlands


adopters seem fairly similar to each other and fairlydifferent from the involuntary adopters and rejecters.Comparing the decision-making process of adopterswith non-adopter, we see that voluntary adopterstake big decisions independent of others and thatthis group of adopters does not take considerabletime for big decisions yet quits differently from theother three groups. Finally, all the respondents an-swered that rules are necessary in daily life, but theadopters with a positive attitude score a little bithigher than the respondents with a neutral or nega-tive attitude. The majority of the respondents an-swered that traditional norms and values are import.

Furthermore, the findings in the activity spider dia-gram indicate that recycling paper, buying energy effi-cient equipment (A label), and use water wisely aremajor activities performed by all the respondents.Remarkable, the avoidance of unaddressed advertising

is notified as not common by the majority of the respon-dents. With regard to the different adopter groups, wesee that the adopters are more sustainable minded thanthe non-adopters. Within the group of non-adopters,rejecters have indicated that they are less sustainableminded on all the analyzed characteristics than the po-tential adopters. In conclusion, the adoption of PV didnot appear to vary between the respondents based ontheir level of environmental consciousness.

Finally, the value spider diagram suggests that theprice of a system is a major issue perceived by allthe respondents. Remarkable, the efficiency is men-tioned as more important for the voluntary and po-tential adopters, thus, for PV respondents who havepositive attitude, while the visual representation ismentioned by the involuntary adopters and rejecters,the group respondents with a neutral or negativeattitude on PV. The integration is more important

Fig. 2 Age, income, andeducation of all respondentsdivided over the different adoptergroups


by the group involuntary adopters and quite differ-ent from the other groups.

Behavioral characteristics

Figure 5 gives an overview of the behavioral char-acteristics, and important aspects for this group ofcharacteristics are particularly asked to adopters ornon-adopters (see BResearch method and datacollection^), and therefore, it was not possible toinclude all adopter groups in this analysis; neverthe-less, we can describe these characteristics. Bothspider diagrams consist of the most import barriersand motives, and more aspects are included in thequestionnaire, but the results indicated that theseaspects have not a role for the adoption or rejectionof PV.

The barrier spider diagram suggests that for thevast majority of non-adopters (potential adoptersand rejecters), the high investment costs of PV isthe most important aspect followed at a large dis-tance with low energy yield. With regard to the

efficiency of a PV system6, we see that potentialadopters have more fear for gaining promised effi-ciency than the group rejecters. In line with ourexpectations, many rejecters indicated that they arenot interested in adopting a PV system.

Concerning motivation of adoption of PV, for themajority of the voluntary adopters, the saving of elec-tricity costs is the most import aspect together with thecosts of a PV system. The possibility to be self-sufficientand the contribution to a better natural environment arealso important motivations for adoption. The visualrepresentation and innovativeness of the system arenot seen as an important aspect. This is against ourexpectations because the visibility of the technologycan function as a status symbol or serve as a symbol tocommunicate a certain identity or value orientation. Anelegant and aesthetically integrated system should be

Fig. 3 Housing type, housingsituated, ownership, and numberof residents of all respondentsdivided over the different adoptergroups

6 The PV technology keeps on developing, and its main compo-nent, the PV cells, is expected to become more efficient inconverting solar energy into electricity (IEA 2010). The perfor-mance ratio, related to the efficiency of a PV system, is alsoexpected to grow in the future (Jahn and Nasse 2004).


used to convince neighbors, friends, and family of thepositive role that PV can play.

Finally, discussion with other adopters to convinceadoption is not an issue; the non-adopters do not feelpressure from their social environment.

Overview of the characteristics

Table 2 gives an overview of the demographic charac-teristics, geographic characteristics, psychographiccharacteristics (cultural beliefs and the importance ofthe different attributes for PVof the different lifestyles),and behavioral characteristics (motives and barriers toadopt PV).

Results show that voluntary adopters are on averagemiddle-aged, highly educated, take big decision inde-pendent of others, and take care of the environment byfor example recycling paper and avoiding the car on aregular basis. The opposite are the rejecters who have on

average a lower income, take big decisions dependenton others, and need also considerable time for big deci-sions. These characteristics help to construct a picture ofwhich kind of people adopt or reject a system; however,we cannot do predictions based on these characteristics.Therefore, we indicate some determining factors whichcorrespond with some psychographic and behavioralcharacteristics. Especially characteristics regarding thecost and benefits of adoption are of importance. In thisresearch, the price of a system is indicated as the mostimportant issue perceived by the different groups7. It isnot only the most important motive to adopt but also themost important barrier to reject (potential adopter andrejecter). The financial arguments for PV are

Fig. 4 Psychographiccharacteristics of all respondentsdivided over the different adoptergroups

7 Since the questionnaire have not asked people about the relativeimportance of the several aspects of this price (purchase, operatingcosts, maintenance costs, and insurance rates), it is impossible todetermine which are perceived as most important.


complicated by amongst others the uncertainty of elec-tricity prices over the 20- to 30-year system life. Manyrespondents felt that prices would inevitably rise, andtherefore, PV could provide a good option and a degreeof self-sufficiency. While self-sufficiency is an impor-tant part of owning a PV system, PV did not provide thelevel of independence that we expected (see Bbehavioralcharacteristics^, only ±30 % of the adopters citing theimportance of self-sufficiency). The motive to contrib-ute to a better natural environment appeared to be moreimportant for the group voluntary adopters than thesymbolic investment. Regarding the non-adopters, wecan conclude that for the potential adopters next to thecosts of a PV system, the efficiency and energy yield areimportant determining factor. This is different for thegroup rejecters; they indicated the visual representationas a barrier for the adoption of PV, while this aspect ispointed out as an important aspect by analyzing thepsychographic characteristic. These findings suggestthat the visual representation is not a determining factorfor this group of adopters. As a barrier, they indicatednext to the costs and the energy yield that they are not

interested. To conclude, PV-specific characteristics aremore important for the analysis of the adoption or rejec-tion process of a PV system than characteristics regard-ing demography and geography. Interestingly, otherthan demographic and geographic characteristics, thedifferent groups exhibit more similar psychographiccharacteristics. These similarities show us howintertwined or entangled self-concept and personal in-terest is with actual and perceived knowledge and com-petence and the actual and reported performance of apractice.

Our results are different in several ways from thosefound in classic diffusion research such as the theory ofBeal and Bohlen (1957) and Rogers (2003). People withthe highest socioeconomic status who are the least guid-ed by others in their decision making do not necessarilyfall into the first groups of adopters (innovators) asdefined by Rogers. In this research, the group of volun-tary adopters is in line with this thought, but the invol-untary adopters not. Also, Rogers’ group of householdsthat never will adopt a system (laggard/non-adopters) isnot in line with our research. This group should have a

Fig. 5 Behavioral characteristicsof all respondents divided overthe different adopter groups


Tab

le2

Characteristicsof

thedifferentlifestyles

Adopters

Non-adopters

Voluntary

Involuntary

Potentialadopters

Rejecters

Dem

ographiccharacteristics

Age

50–59

30–49

40–59

40–59

Income

Onaveragehigher

Onaveragehigher

Onaveragelower

Onaveragelower

Educatio

nHigh

High-middle

Middle

Middle-low

Gender

Sim

ilar

Sim

ilar

Sim

ilar

Sim

ilar

Geographiccharacteristics

Hom

e-ow

ner

Me

Rental

Me

Rental

Housing

situation

Morein

avillage

Morein

avillage

Morein

acity

Morein

acity

Housing

type

Middleof

row/detached

Middleof

row/sem

i-detached

Middleof

row/sem

i-detached

Middleof

row/apartment

Num

berof

residents

23or

4Similar

Sim

ilar

Psychographiccharacteristics(opinion)

Clim

atechange

Moreim

portant

Lessim

portant

Moreim

portant

Lessim

portant

Traditio

naln

ormsandvalues

Important

Important

Important

Important

Rules

Necessary

indaily

life(100

%)

Necessary

indaily

life(98%)

Necessary

indaily

life(95%)

Necessary

indaily

life(96%)

Tim

eneeded

formakingbigdecisions

Lessconsiderable

Considerable

Considerable

Considerable

Taking

bigdecisions(in)dependento

fothers

Independent

Moredependent

Moredependent

Moredependent

Psychographiccharacteristics(activity

)

Recyclin

gof

paper

(Alm

ost)always

(Alm

ost)always

(Alm

ost)always

(Alm

ost)always

Avoidance

ofunaddressedadvertising

Occasional

Occasional

Occasionaltoalesser

extent

Occasionaltoalesser

extent

Energy-efficientequipment

(Alm

ost)always

(Alm

ost)always

(Alm

ost)always

Regularly

Avoidance

ofcaruse

Regularly

Regularly

Regularly

toalesser

extent

Regularly

toalesser

extent

Water

conservatio

n(A

lmost)always

(Alm

ost)always

(Alm

ost)always

(Alm

ost)always

Psychographiccharacteristics(value)

Efficiency—

amount

ofenergy

Moreim

portant

Lessim

portant

Moreim

portant

Lessim

portant

Integration

Not

important

Moreim

portant

Lessim

portant

Lessim

portant

Lifetim

eof

thesystem

Lessim

portant

Not

important

Lessim

portant

Lessim

portant

Price

Mostimportant

Mostimportant

Mostimportant

Mostimportant

Visualrepresentation

Not

important

Important

Not

important

Important

Behavioralcharacteristics(m

otive)

Costsefficientsystem

Mostimportant

––

–


Tab

le2

(contin

ued)

Adopters

Non-adopters

Voluntary

Involuntary

Potentialadopters

Rejecters

Savingson

electricity

bill

Mostimportant

––

–

Qualityof

thesystem

Lessim

portant

––

–


Lessim

portant

––

–

Self-sufficient

Moreim

portant

––

–

Innovativ

esystem

Lessim

portant

––

–

Easyto

install

Lessim

portant

––

–

Lessenvironm

entalimpact

Moreim

portant

––

–

Behavioralcharacteristics(barrier)

Investmentistoohigh

––

Mostimportant

Mostimportant

Energyyieldistoolow

––

Moreim

portant

Moreim

portant


––

Lessim

portant

Lessim

portant

Difficultto

install

––

Lessim

portant

Lessim

portant

Fearof

gainingprom

ised

efficiency

––

Moreim

portant

Lessim

portant

Fear

ofsubsidyadjustments

––

Lessim

portant

Lessim

portant

Not

interested

––

Lessim

portant

Moreim

portant


low socioeconomic status which are set to belong to thegroup of households that is least guided by others intheir decision-making. In this research, the group ofrejecters is more guided by others in their decisionsand has middle to low socioeconomic status, comparedto Rogers’ middle groups (early and late majority).

Reflection

To test whether the introduced segmentation models canbe used as a good segmentation for technological inno-vations, we analyze how the segmentation model metthe seven criteria of Gankema and Wedel (identifiable,accessible, size, heterogeneous, stable, homogeneousresponse, and conducive of steering). The typologyscored well on most of the criteria. It is clear to whichsegment somebody belongs, the segments were easy todistinguish (heterogeneity) and stable, and the groupsare big enough and gave in general homogenous re-sponses in the questionnaire. The criteria which wasmet least, was the criteria of accessible, which says thatpeople should have the opportunity to move in andmove out a certain group. It is obvious that it is notlogical to move from the group voluntary adopter to thegroup potential adopters, except when you moved toanother dwelling. However, to a certain extent, it ispossible to move in another group. Potential adopters,for example, can become adopters of the technology,and rejecters of the technology can also become poten-tial adopters which in turn can become adopters of thetechnology, see Fig. 6. Moreover, it is also possible tomove from involuntary adopters to voluntary adopters.Involuntary adopters, who do not decide the purchase of

the system themselves, can be very positive of thesystem and can decide to adopt more panels and in thisway become a voluntary adopter. The other way aroundis less obvious (from voluntary to involuntary adopters)as the purchase decision of the panels cannot be changedfrom themselves to another person who decided.

The last criterion, conducive of steering, is more orless connected to the accessibility of the typology. Everygroup should give a description about preferred policyoptions and/or aspects where people in the group strivefor. A favorable grant, for example, can have an influ-ence on the behavior of people within a certain group.The grant can stimulate potential adopters to becomevoluntary adopters.

Our segmentation model is constructed on surveyresearch and on scientific insights and offers opportuni-ties for analyzing, exploring, and visualizing beliefs andperspectives of people who are in the adoption processof PV. It gives more specific insight in the differentbeliefs and perceptions of the adoption of PV. Thesegmentation method can be used to classify, interpret,and analyze these different beliefs and perspectives. Inthis way, they can be used to analyze the response inorder to contribute to the Dutch energy system in theyears ahead and the future social acceptance of differenttechnological innovations. By doing this, insights can beprovided how the government’s policy or service canalign the needs of the customer (citizen) as well as howsuppliers of this technology can optimize their productbased on identified consumer beliefs and preferences.

Furthermore, a first attempt is made to includeinnovation-specific characteristics during the segmenta-tion of a larger populace. Characteristics (more in par-ticular psychographic and behavioral) related to the cost

Fig. 6 Accessibility andinfluentially of the differentgroups


and the benefits are taken into account. The benefits ofan innovation obviously refer to the positive consequences(e.g., environmental benefits), while the costs refer to thenegative consequences (e.g., financial uncertainty). In thisway, our segmentation model can be distinguished fromthe segmentation models we compared from the differentresearch centers, as their segmentation is not product orservice specific.

Conclusion and discussion

The overall purpose of this paper was to introduce asegmentation model which can be used to determineempirically groups of PV technology adopters. The useof questions about beliefs and attribute preferenceshelped to group people into different groups and comparethe choices for such groups. The empirical analysis, basedon a questionnaire among 817 Dutch households, result-ed in new introduced segmentation model. The modelconsists of four segments which is determined by twoaxes: the view citizens have on the technology (positiveversus neutral or negative) and the decision-making pro-cess of major technological innovations. The four seg-ments are voluntary adopters, involuntary adopters, po-tential adopters, and rejecters. The groups (segments) metthe set criteria for good segmentation and differ from eachother with respect to the demographical, geographical,physiographical, and behavioral characteristics. We splitour results in non-determining and determining factors toexplain how the groups differ from each other. Regardingthe determining factors, the costs of a PV system areincluded as a benefit of having PV for voluntary adopterswhich obviously refer to the positive consequences ofhaving a system (e.g., self-sufficiency and environmentalbenefits), while for non-adopters, the benefits of PVwhich also refer to the positive consequences of havinga system do not outweigh the costs of a system whichrefer to the negative consequences (e.g., price and finan-cial uncertainty). Regarding the non-determining factors,voluntary adopters are on average middle-aged, highlyeducated, take big decision independent of others, andtake care of the environment by for example recyclingpaper and avoiding the car on a regularly basis. Theopposite are the rejecters who have on average a lowerincome, take big decisions dependent on others, and needalso considerable time for big decisions.

Because of the groups dynamic nature, the interpre-tation of beliefs and attribute preferences may change

over time, as well as the distribution between the differ-ent groups (see also Fig. 6). The distribution betweenthe groups change due to events (e.g., air pollution,global warming), developments (e.g., price of the tech-nology or technical performance), and occurrences (pos-sibly catalyzed by people or the media). For example,the symbolism of PVmay change over time, as has beenseen with perceptions of the automobile (Sachs 1983).An accumulation of these events, developments, andoccurrences may lead to a changing dominant group(e.g., voluntary adopters instead of rejecters), and thesupport for a given strategymay also change. Hence, thesegmentation model, including the used method, remainsthe same, but results may be different as different scoreson indicators (characteristics) may be found. Although itseems that these characteristics play a role in segmenta-tion, it is not clear what role they exactly play. Onepossibility is that they do have a direct influence (demo-graphic and geographic characteristics) on people’s per-spectives; another possibility is that they have an influ-ence on people’s taste, preferences, and values, which areon their turn determining for one’s group. Psychographiccharacteristics may also increase our understanding ofgroup change as change may then be related to changingdemographic distributions and population characteristics.These changes and the effect on adoption should besubject of further study.

After this analysis, it becomes clear that for thebreakthrough of technologies, the innovation-specificcharacteristics are more important than characteristicsregarding demography and geography. However, thecost and benefits we used in this analysis is a firstattempt and can be extended. For example, we do notask people about the relative importance of the severalaspects of the price, for example, purchase price, oper-ating costs, maintenance costs, and insurance rates.Also, the expected future price of electricity has to betaken into account. Including these, aspects make itpossible to determine which costs are perceived as mostimportant. Furthermore, the social influence and theeffort people must do to adopt a system are not thor-oughly considered in this research. Thus, further re-search remains necessary on how to further integratethis cost-benefit dimensions and the social influence insegmentation analysis and on how and when to includethese aspects.

Insights regarding both adopters (voluntary and in-voluntary) and non-adopters (potential adopters andrejecters) of PVare provided in this research. Four major


groups of characteristics are taken into account: thedemographical characteristics, the geographic character-istics, the physiographical characteristics, and the be-havioral characteristics. Overall, the link between thecharacteristics/attitudes we see in this research and theactual adoption behavior is found not to be strongenough. Therefore, it would be beneficial to study thefourth group, behavioral characteristics, in more detailto better understand their motivations to (not) purchase asystem. Not only the perceived advantage a PV systemhave but also the complexity people perceive are twoexample concepts which have to be taken into accountto help gauge people perception of PV which shape theadoption process. By doing this, insights can be obtain-ed about useful government’s policies fitting in with theneeds of the customer (citizen) and how suppliers of thistechnology can optimize their product based on identi-fied consumer preferences and frames.

It is important to notice that the case analyzed in thispaper deals with a particular sustainable energy technol-ogy. The behavioral response related to this technologymight be quite specific, and response related to othersustainable energy technologies may be quite different.Also, the composition of the respondents may be quitedifferent. In the case of PV, we expect that the potentialadopters may become less common in the near future,particularly as the government and municipalities en-courage installation of PV in social housing or otherlarge development. As such, the potential adoptersmay have markedly different socioeconomic, attitudinaland value characteristics, and hence, behavioral re-sponses. Nevertheless, describing segments in a system-atic way allows us to compare different cases and toidentify typical responses that may be associated withsuccess or failure. This provides insights regarding theadoption of a technology and provides insights forpolicymakers especially when considering the imple-mentation of market formation policies for sustainabletechnologies. Further research is therefore necessary toexpand the empirical cases not only to different technol-ogies but also to different sectors and countries. In orderto generalize research findings, it is important to use thesame segments and take the different characteristicgroups into account.

It is possible that a small number of the respondentswho do not adopt a system have for example not enoughknowledge to fill in the questionnaire, because they donot want it, cannot afford it, or do not understand it.These people are also included in this research and

impossible to omit. This is seen as a limitation to thisresearch.

Appendix

Valuebox model of NFO TrendBox

TrendBox is a strategic market research agency, special-izing in qualitative and quantitative research on brands,people, and their motives. In 1990, TrendBox started theLife & Living project, an ongoing study where the atti-tudes, behavior, and mentality of the Dutch are trackedover time. Because of the continuous nature of Life &Living, NFO TrendBox is able to identify the status quo,recognize, and analyze coming trends and translate thefindings to the future. TrendBox distinguish six clustersof segments (order and decency, purposeful adventure,center, uncomplicated beneficiary, sober philosophy, spir-itual) in which social groups of the Dutch population areclassified, called the valuebox model (de Wit 2003;Meijer et al. 2008; Trendbox 2011).

Mentality model of Motivaction

Motivaction, a Dutch research centre, developed thementality test which is a value and lifestyle researchmethod and focuses on marketing and policy questions.Results are applicable to, e.g., sustainability issues, mo-bility, media, and politics. Within the typology, eightsocial environments are distinguished (traditional citi-zens, comfort oriented, modern citizens, new conserva-tives, cosmopolitans, upward mobiles, postmaterialists,postmodern hedonists) which differ in terms of status(low importance, middle importance, and high impor-tance) and values (traditional, based on conservation;modern, based upon possession and spoil; or postmodern,based on self-development and experience) (Motivaction2010a, b).

Dutch society is highly individualized, and thereis a wide variety of lifestyles; citizens are also moremature and become more critical. How can you as apolicymaker, advisor, or manager keep in touch withwhat people moves? In order to deal with this question,Motivaction introduced four styles of citizenships(dutiful, responsible, pragmatic, and outsiders) incollaboration with the Commission Future GovernmentCommunication and Scientific Council for GovernmentPolicy. The citizenship styles are based on the Mentality


test of Motivaction which is conducted since 1997in the Netherlands among Dutch people between15 and 80 years old. The citizenship styles do notonly provide insights into the opinions of peopleand the activities they undertake but they also giveinsight into the motivations, the needs they have,how they can be involved in the public domain,and how they can be most effectively addressed(Motivaction 2010a). So, the different styles of citi-zenship can represent the attitude toward governmentand politics.

Mosaic model of Experian

MOSAIC is a geo-demographic segmentation systemdeveloped by Experian and marketed in over 20 coun-tries worldwide. In the Netherlands, Mosaic has divided16 million Dutch people into 10 groups and is classifiedinto 44 segments (different types of consumers). The 10groups are free spirited, the developed urban dweller,go-getter, dynamic families, modal citizens, successfulfamilies, traditionalists, rural family life, well-off peo-ple, and pension beneficiary. This classification hasbeen based on sociodemographic and socioeconomicdata, lifestyle, preferences, and (buying) behavior(Experian 2012; Meijer et al. 2008).

WIN model of TNS/NIPO

The WIN model is a value and sociodemographiccharacteristics based segmentation of the Dutch popu-lation. The different values that people find importantin their lives seem to be related to different ways oflife, housing, dress, think, consume, and vote. Scoreson a vertical (focused on others) and horizontal (ex-ploring possibilities) axes are used to determine theclassification of segments (Hessing and Reuling 2003;TNS-NIPO 2011). The model distinguishes eightgroups in society, which are very different in termsof lifestyle, attitudes, motivations, and behavior. Thegroups are engaged, care takers, conservatives, hedo-nists, luxury seekers, professionals, broad minded, andbalanced (TNS-NIPO 2011).

Censydiam model of Synovate

Censydiam model is based on consumer motivationstudies, including decision to buy. Motivations are

fundamental human desires that drive behavior. Themodel is a basis for systematical understanding of peo-ple motivation in their connection with brand position-ing and communications. The model is built around twomain axes: personal dimension (how they feel in relationto themselves) and social dimension (how people feel inrelation to other people). Around these axes, life valuesare placed. The segments are vitality, enjoyment, con-viviality, belonging, security, control, recognitions, andpower (Censydiam 2011).

BSR model of SmartAgent Company

SmartAgent is a perception and consultancy center.SmartAgent gains experiences of people using thesocial-psychological model, Brand StrategyResearch (BSR) model, which is applied in quali-tative and quantitative research. The BSR modelexplores and structures the underlying values,needs, and motivations of people within a particu-lar domain. The model is visualized by two be-havioral dimensions: the sociological (x-as) andpsychological dimension (y-as). In this way, fourquadrants emerge, in other words, four experiencesfrom which people think and act (Coppens andOosterlynck 2008; Smartagent 2011).

In cooperation with MarketResponse-Amersfoort(market research center) and Kolpron Consultans-Rotterdam (focuses on market research and advicein the areas of the built environment), two livingexperiences are investigated by SmartAgent. Thestudies, conducted in 1998 and 2000, form thebasis for a subdivision into six experience profiles(living together, withdrawals, dynamic individual-ist, anchored, quiet luxury, and unattached). Eachprofile describes a social group with a similar setof values and behaviors related to housing behav-ior and housing preferences. People score on allclusters in a given ratio. Based on such profiles,the preference for specific living environments canbe established. In this way, there is a direct relationbetween lifestyle and living environment (Coppensand Oosterlynck 2008; Smartagent 2011).

Open Access This article is distributed under the terms of theCreative Commons Attribution License which permits any use,distribution, and reproduction in any medium, provided the orig-inal author(s) and the source are credited.


References

Abell, D. F., & Hammond, J. S. (1979). Strategic market planning,problems and anallytical approaches. Prentice-Hall.

Axsen J, TyreeHageman J, & Lentz A (2012). Lifestyle practicesand pro-environmental technology. Ecological Economics(82), 64-74.

Beal, G. M., & Bohlen, J. M. (1957). The diffusion process.Special Report No. 18 (Vol. (reprinted 1-67, 10-71, 5-75,and 11-81)). Ames, Iowa: Cooperative Extension Service,lowa State University of Science and Technology.

Beal, G.M., Rogers, E. M., & Bohlen, J. M. (1957). Validity of theconcept of stages in the adoption process. Rural Sociology,22(2), 166–168.

Beane, T. P., & Ennis, D. M. (1987). Market segmentation: areview. European Journal of Marketing, 21(5), 20–42.

Censydiam (2011). The Censydiam model explained. http://www.censydiam.com/modelexplained.asp. Accessed October 4 2011.

Coppens, A., & Oosterlynck, S. (2008). Leefstijlbenaderingen enruimtelijk gedrag. Een literatuurstudie naar uitdagingenvoor het ruimtelijk beleid. Rapport Globale MaatschappelijkeDynamiek I - Leefstijlen. Heverlee: Steunpunt Ruimte enWonen.

de Wit, M. (2003). Segmentatie: Hoe doe je dat? Amsterdam:Kluwer.

Doornbos, L. (2004). Vraaggestuurd werken en segmentatie.Veranderd ondernemen in de non-profisector. Maarssen:Elsevier Gezondheidszorg.

Experian (2012). Mosaic Huishouden: Van database naarklantgesegmenteerde communicatie Den Haag: ExperianNederland B.V.

Faiers, A., & Neame, C. (2006). Consumer attitudes towards domes-tic solar power systems. Energy Policy, 34(14), 1797–1806.

Feather, N. T. (1975). Values in Education and Society. New York:Free Press.

Gankema, H.G. J.,&Wedel,M. (1992).Marktsegmentatieonderzoekin theorie en praktijk. In Recente Ontwikkelingen in hetMarktonderzoek (Vol. Jaarboek 1992-’93, pp. 207-222).Haarlem: De Vrieseborch.

Geest, T., Jansen, J., Mogulkoç, E., Vries, P., & Vries, S. (2008).Segmentation and E- Government; a literature review.Universiteit Twente & Telematica Instituut.

Hawkins D, Mothersbaugh D, & Best R (2007). ConsumerBehavior: Building Marketing Strategy (10th ed.).

Heitgerd, J. L., & Lee, C. V. (2003). A new look at neighborhoodsnear National Priorities List sites. Social Science &Medicine,57(6), 1117–1126.

Hessing, E., & Reuling, A. M. H. (2003). Waarden in Nederland:segmentatie van doelgroepen. In MarktOnderzoekAssociatiesJaarboek 2003 (pp. 163- 175).

IEA. (2010). Technology Roadmap. Solar photovoltaic energy.Paris: International Energy Agency.

Jager, W. (2006). Stimulating the diffusion of photovoltaic sys-tems: a behavioural perspective. Energy Policy, 34, 1935–1943.

Jahn, U., & Nasse, W. (2004). Operational Performance of Grid-connected PV Systems on Buildings in Germany. Progress inPhotovoltaics: Research and applications, 12(6), 441–448.

Kahle, L. R., Beatty, S. E., & Homer, P. (1986). Alternative mea-surement approaches to consumer values: the list of values(LOV) and values and life style (VALS). Journal of ConsumerResearch, 13(3), 405–409.

Labay, D. G., & Kinnear, T. C. (1981). Exploring the consumerdecision process in the adoption of solar energy systems.Journal of Consumer Research, 8(3), 271–278.

Maslow, A. H. (1954). Motivation and Personality. New York:Harper.

Meijer, E.,Waning, E., &Dubbeldam, B. (2008).Klantsegmentatie:Nederlandse openbare bibliotheken. Rotterdam: Cultuuronderzoeken, de kunst van cultuurmarketing.

Mitchell, A. (1983). Nine American lifestyles: Who we are andwhere we are going. New York, NY: MacMillan.

Motivaction (2010a). Mentality™-model. Accessed February 72010.

Motivaction (2010b). Burgerschapsstijlen. Accessed February 72010.

NetMBA (2010). Market Segmentation. http://www.netmba.com/marketing/market/segmentation/ Accessed 26 October 2012.

Pedersen, L. (2000). The dynamics of green consumption: a matterof visibility? Journal of Environmental Policy & Planning2(3), 193–210.

Reisman, D., Glazer, N., & Denney, R. (1950). The Lonely Crowd.New Haven: Yale University Press.

Rogers, E. M. (2003). Diffusion of innovations (Fifthth ed.). NewYork: The Free Press.

Rokeach, M. (1973). The Nature of Human Values. New York:Free Press.

Sachs, W. (1983). Are energy-intensive life-images fading? Thecultural meaning of the automobile in transition. Journal ofEconomic Psychology, 3(4), 347–365.

Schiffman, L. G., & Kanuk, L. L. (2010). Consumer Behavior -Global Edition (tenthth ed.). New Jersey: Prentice Hall.

Smartagent (2011). Het BSR-model. http://www.smartagent.nl/.Accessed October 4 2011.

TNS-NIPO (2011). Het WIN- model. http://www.tns-nipo.com/expertise/business-solutions/win2/. Accessed october 4 2011.

Trendbox (2011). Trends: life & living. http://trendbox.nl/.Accessed October 4 2011.

Wedel, M., & Kamakura, W. A. (2000). Market segmentation:conceptual and methodological foundations. Dordrecht:Kluwer.


http://www.censydiam.com/modelexplained.asp

http://www.censydiam.com/modelexplained.asp

http://www.netmba.com/marketing/market/segmentation/

http://www.netmba.com/marketing/market/segmentation/

http://www.smartagent.nl/

http://www.tns-nipo.com/expertise/business-solutions/win2/

http://www.tns-nipo.com/expertise/business-solutions/win2/

http://trendbox.nl/

Documents

A segmentation analysis: the case of photovoltaic in …...analysis into the adoption and non-adoption of solar PV in Dutch households. It is based on a survey under 817 households